5.10-dars: Ko'p grafik (subplots)

5-QISM — VIZUALIZATSIYA · 10-dars

1. Kirish va motivatsiya

Ba'zan bitta grafik yetmaydi — bir nechta grafikni yonma-yon ko'rsatish kerak: 4 xususiyatning histogrami, bir necha yilning trendi, yoki taqqoslash uchun. Subplots (kichik grafiklar) — bitta figurada bir nechta grafik (panel to'ri): fig, axes = plt.subplots(2, 2) — 2×2 = 4 panel, har biriga alohida chizasiz. 5.1-darsda fig, ax = plt.subplots() — bitta panel; endi bir nechta. Data Scientist EDA (ma'lumot o'rganish) paytida bir nechta xususiyatni bir qarashda ko'rish uchun subplots ishlatadi. Nega muhim? (1) Taqqoslash — grafiklar yonma-yon (bir figurada); (2) EDA — ko'p xususiyat bir qarashda; (3) Tuzilma — panel to'ri (tartibli). Bu dars subplots'ni o'rgatadi — bir figurada ko'p grafik.

Subplots (ko'p grafik) — bir figurada bir necha panel: plt.subplots(qator, ustun) (panel to'ri — fig, axes qaytaradi; 5.1 figure/axes), axes[i] yoki axes[i, j] (har panel — alohida ax), har ax'da (ax.plot, ax.set_title — 5.1), figsize= (umumiy o'lcham), tight_layout() (panel orasi joy — to'silmaslik), sharex/sharey (umumiy o'q — taqqoslash). Foydalanish: taqqoslash, EDA, panel to'ri. Bu 5.1 (figure/axes), 5.7 (Seaborn ax=) bilan bog'liq. Subplots — bir figurada ko'p grafik. Panel. Taqqoslash.

Real vaziyat. Data Scientist uy ma'lumotida 4 xususiyat (maydon, narx, yosh, xonalar) taqsimotini ko'rmoqchi. Bitta grafik: 4 alohida rasm (fayl — chalkash; taqqoslash qiyin). Subplots (fig, axes = plt.subplots(2, 2, figsize=(10, 8))): 2×2 = 4 panel bir figurada; axes[0, 0].hist(maydon), axes[0, 1].hist(narx), ... (har panel — bir xususiyat); fig.tight_layout() (panel orasi joy). Bir qarashda 4 taqsimot (maydon normal, narx o'ngga cho'zilgan (5.5 chapga/o'ngga), yosh bir tekis, xonalar diskret) — taqqoslash (bir figurada; EDA). Data Scientist ko'p xususiyatni bir qarashda ko'rdi (subplots — panel to'ri). Subplots — bir figurada ko'p grafik.

Bu darsda subplots'ni o'rganamiz.

Bu darsda:

• subplots (panel to'ri)
• axes indekslash (har panel)
• figsize, tight_layout (o'lcham, joy)
• sharex/sharey (umumiy o'q)
• Seaborn subplots (ax=)
• Subplots amaliyoti
• Subplots tuzoqlari
• Amaliy: subplots modeli

ℹ Misollar real matplotlib bilan (Agg — grafik xususiyatlari matn bilan tekshiriladi) ishlaydi.

2. Nazariya — chuqur tushuntirish

2.1. subplots (panel to'ri)

Panel to'ri:

import matplotlib.pyplot as plt

# subplots(qator, ustun) — panel to'ri
fig, axes = plt.subplots(2, 2)   # 2×2 = 4 panel

# fig — butun figura; axes — panellar massivi

subplots (panel to'ri) — bir figurada ko'p panel: plt.subplots(qator, ustun) — panel to'ri yaratadi (fig, axes qaytaradi; fig — butun figura, axes — panellar massivi; 5.1 figure/axes). Sabab: bitta plt.subplots() 5.1-bob — bir panel; ko'p grafik uchun to'r ((2, 2) — 2 qator × 2 ustun = 4 panel; (1, 3) — 1 qator × 3 = yonma-yon 3). fig (butun figura — saqlash, umumiy sarlavha), axes (panellar — har biriga chizish). Panel to'ri ((qator, ustun)), fig (butun), axes (panellar). subplots — panel to'ri (bir figurada ko'p). Panel. To'r.

2.2. axes indekslash (har panel)

Har panelga:

import matplotlib.pyplot as plt

fig, axes = plt.subplots(2, 2)   # 2D massiv

# har panel — axes[qator, ustun]
axes[0, 0].plot(...)   # yuqori chap
axes[0, 1].hist(...)   # yuqori o'ng
axes[1, 0].scatter(...)   # past chap

# 1 qator — axes[i] (1D)
fig2, axes2 = plt.subplots(1, 3)
axes2[0].plot(...)

axes indekslash (har panel) — massiv: subplots(qator, ustun) axes — massiv (2D — axes[qator, ustun]; 1 qator/ustun — 1D axes[i]); har panelga alohida chizish (axes[0, 0].plot, axes[0, 1].hist — 5.1 ax metodlari). Sabab: har panel — mustaqil ax (o'z grafigi, sarlavha, o'qi — 5.1); indeks bilan tanlash ([0,0] yuqori chap, [1,1] past o'ng); ax metodlari (plot, hist, scatter, set_title — 5.1-5.6). 2D ((2,2) — axes[i,j]), 1D ((1,3) yoki (3,1) — axes[i]), axes.flat (tekis — sikl uchun). axes indekslash — axes[i,j] (har panel alohida). Massiv. Panel.

2.3. figsize, tight_layout (o'lcham, joy)

figsize, tight_layout — o'lcham va joy: figsize=(kenglik, balandlik) (butun figura o'lchami — dyuym; ko'p panel — katta; subplots(2, 2, figsize=(10, 8))), fig.tight_layout() (panel orasi joy — sarlavha/yorliq to'silmasligi; avtomatik joylashtirish). Sabab: ko'p panel — o'lcham (kichik figura — panellar siqiladi; katta — o'qiladi); tight_layout — panellar to'sadi (sarlavha/o'q yorliqlari ustma-ust; tight_layout — joy qo'shadi). figsize (umumiy o'lcham — panel soniga qarab), tight_layout (to'silish — joy), constrained_layout=True (yangi — avtomatik joy). figsize (o'lcham), tight_layout (joy). figsize/tight_layout — o'lcham va joy (siqilish/to'silish). O'lcham. Joy.

2.4. sharex/sharey (umumiy o'q)

sharex/sharey (umumiy o'q) — taqqoslash: sharex=True/sharey=True (panellar bir xil o'q — x yoki y; taqqoslash oson). Sabab: panellar turli o'q (avtomatik — har panel o'z masshtabi; taqqoslash noto'g'ri — bir xil balandlik turli qiymat); sharey=True — bir xil y (taqqoslash — balandlik to'g'ri); "aldamas masshtab" (5.13 — yolg'on grafik). sharex (bir xil x — vaqt qatori), sharey (bir xil y — taqqoslash; muhim). sharex/sharey — umumiy o'q (taqqoslash; masshtab). Umumiy. Taqqoslash.

2.5. Seaborn subplots (ax=)

Seaborn subplots (ax=) — birga: Seaborn 5.7-bob grafiklari ax= parametr (subplots panelga chizish; sns.histplot(..., ax=axes[0, 0])); Seaborn (chizadi) + Matplotlib subplots (tuzilma — 5.7 integratsiya). Sabab: Seaborn axes-daraja grafiklari (histplot, scatterplot, boxplot — ax=; panelga); subplots (Matplotlib — panel to'ri) + Seaborn (har panel — ax=); figure-daraja (relplot, displot, catplot — o'z subplots'i; ax= yo'q — chalkash). ax=axes[i, j] (Seaborn panelga), axes-daraja (histplot, boxplot — ax=), figure-daraja (displot — o'zi subplots; aralashtirma). Seaborn subplots — ax= (axes-daraja; Matplotlib to'r + Seaborn chizish). Seaborn. ax=.

2.6. Subplots amaliyoti

Subplots amaliyoti: fig, axes = plt.subplots(qator, ustun, figsize=(k, b)) (panel to'ri + o'lcham); axes[i, j].plot/hist/... (har panel — 5.1-5.6); axes[i, j].set_title(...) (har panel sarlavha); fig.suptitle(...) (butun figura sarlavha); fig.tight_layout() (joy — to'silish); sharey=True (taqqoslash — umumiy y); fig.savefig(...) (butun figura — 5.1); Seaborn (ax=axes[i, j] — 5.7). Tuzoqlar: figsize kichik (siqilish), tight_layout unut (to'silish), axes 1D/2D (indeks xato), turli o'q (taqqoslash — sharey), figure-daraja (ax= yo'q). Amaliyot — subplots, axes, tight_layout, sharey. Panel. Taqqoslash.

2.7. Subplots tuzoqlari

Subplots asosiy tuzoqlari: figsize kichik (ko'p panel — kichik figura (figsize yo'q — standart 6×4); panellar siqiladi (o'q/yorliq o'qib bo'lmas); figsize katta — panel soniga qarab (2×2 — ~10×8)); tight_layout unut (panellar to'sadi (sarlavha/o'q yorliqlari ustma-ust); fig.tight_layout() yoki constrained_layout=True — joy); axes 1D vs 2D (subplots(2, 2) — 2D (axes[i, j]); subplots(1, 3) — 1D (axes[i]); subplots(1, 1) — skalyar (axes — massiv emas; [i] xato); indeks turi e'tibor); turli o'q taqqoslash (panellar turli masshtab (avtomatik — har panel o'z min-max); taqqoslash noto'g'ri (bir xil balandlik — turli qiymat; 5.13 aldash); sharey=True — umumiy y); figure-daraja Seaborn (sns.displot/relplot/catplot — o'z subplots'i (figure-daraja; ax= yo'q); Matplotlib subplots bilan aralashtirma (xato); axes-daraja (histplot, boxplot — ax=)); bo'sh panel (subplots(2, 3) — 6 panel, 5 grafik; 6-panel bo'sh (o'chir — fig.delaxes(axes[1, 2]))); suptitle vs title (fig.suptitle — butun figura; ax.set_title — bir panel; chalkashtirma); savefig figura (fig.savefig — butun figura; ax — bir panel emas). Sabab: subplots tuzilma/o'lcham/indeks nozik (siqilish, to'silish, indeks, masshtab). Yechim: figsize katta, tight_layout, indeks turi, sharey. Tuzoqlar — figsize, tight_layout, indeks, taqqoslash.

2.8. Subplots — bir figurada ko'p grafik

Subplots asosiy g'oyasi — bir figurada ko'p grafik: bitta panel (5.1 — fig, ax) yetmaganda, panel to'ri (plt.subplots(qator, ustun) — bir figurada ko'p grafik; taqqoslash, EDA). fig, axes = plt.subplots(2, 2, figsize=(10, 8)) (panel to'ri + o'lcham), axes[i, j] (har panel — alohida ax; plot/hist/scatter — 5.1-5.6; set_title), figsize (umumiy o'lcham — siqilish), tight_layout (joy — to'silish), sharex/sharey (umumiy o'q — taqqoslash), suptitle (butun figura sarlavha), Seaborn (ax=axes[i, j] — axes-daraja; 5.7). Foydalanish: taqqoslash (grafiklar yonma-yon — bir figurada), EDA (ko'p xususiyat bir qarashda — 9-qism), tuzilma (panel to'ri — tartibli). Tuzoqlar: figsize kichik (siqilish), tight_layout (to'silish), axes 1D/2D (indeks), turli o'q (taqqoslash — sharey), figure-daraja Seaborn (ax= yo'q). Bu 5.1 (figure/axes) ustida va 5.7 (Seaborn ax=), 9-qism (EDA) bilan. Subplots — bir figurada ko'p grafik (taqqoslash, EDA — panel to'ri). Panel. To'r. Taqqoslash.

3. Tez ma'lumotnoma

import matplotlib.pyplot as plt

# PANEL TO'RI:
fig, axes = plt.subplots(2, 2, figsize=(10, 8))   # 2×2 = 4 panel
#   fig — butun figura · axes — panellar massivi

# HAR PANEL (2D — axes[qator, ustun]):
axes[0, 0].hist(maydon);  axes[0, 0].set_title("Maydon")
axes[0, 1].hist(narx);    axes[0, 1].set_title("Narx")
axes[1, 0].scatter(x, y)
axes[1, 1].plot(t, v)

# 1 QATOR (1D — axes[i]):
fig2, axes2 = plt.subplots(1, 3)
axes2[0].plot(...)

# JOY / O'LCHAM / TAQQOSLASH:
fig.suptitle("Umumiy sarlavha")   # butun figura
fig.tight_layout()                 # panel orasi joy (to'silmaslik)
plt.subplots(2, 2, sharey=True)    # umumiy y (taqqoslash)

# SEABORN (axes-daraja — ax=):
sns.histplot(data, ax=axes[0, 0])   # panelga
#   figure-daraja (displot/relplot) — ax= yo'q (o'z subplots'i)

fig.savefig("panellar.png")   # butun figura
QOIDA: figsize katta · tight_layout · indeks 1D/2D · sharey (taqqoslash)

Subplots xulosasi

subplots(qator, ustun) — panel to'ri (fig, axes)
axes[i, j] — har panel (alohida ax; plot/hist/set_title)
figsize/tight_layout — o'lcham va joy (siqilish/to'silish)
sharey — umumiy o'q (taqqoslash — masshtab)
Seaborn — ax= (axes-daraja; figure-daraja emas)

4. Batafsil misollar

Misollar real matplotlib bilan (Agg — grafik xususiyatlari matn bilan tekshiriladi) ishlaydi.

Misol 1 — 2×2 panel to'ri

"""2x2 subplots (real matplotlib)."""

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np


def main() -> None:
    np.random.seed(0)

    print("=== 1. subplots(2, 2) ===")
    fig, axes = plt.subplots(2, 2, figsize=(10, 8))
    print(f"  panellar shakli: {axes.shape} (2x2 = {axes.size})")

    print("\n=== 2. Har panelga chizish ===")
    axes[0, 0].hist(np.random.normal(0, 1, 100))
    axes[0, 1].plot([1, 2, 3], [2, 4, 6])
    axes[1, 0].scatter([1, 2, 3], [3, 1, 2])
    axes[1, 1].bar(["a", "b"], [3, 5])
    print("  4 panel: hist, plot, scatter, bar")

    print("\n=== 3. Har panel sarlavhasi ===")
    axes[0, 0].set_title("Histogram")
    axes[0, 1].set_title("Chiziq")
    print(f"  [0,0]: {axes[0, 0].get_title()}")

    print("\n=== 4. Umumiy sarlavha va joy ===")
    fig.suptitle("4 grafik")
    fig.tight_layout()
    print(f"  suptitle: {fig._suptitle.get_text()}")
    plt.close(fig)
    print("  ⭐ subplots — panel to'ri (4 grafik)")


if __name__ == "__main__":
    main()

Natijaning muhim qismi:

=== 1. subplots(2, 2) ===
  panellar shakli: (2, 2) (2x2 = 4)

=== 2. Har panelga chizish ===
  4 panel: hist, plot, scatter, bar

=== 3. Har panel sarlavhasi ===
  [0,0]: Histogram

=== 4. Umumiy sarlavha va joy ===
  suptitle: 4 grafik
  ⭐ subplots — panel to'ri (4 grafik)

Nima ko'rsatdi: 2.1, 2.2-bo'limlar.

Misol 2 — 1 qator (1D indeks)

"""1 qator subplots: 1D indeks (real matplotlib)."""

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np


def main() -> None:
    np.random.seed(0)

    print("=== 1. subplots(1, 3) — 1D ===")
    fig, axes = plt.subplots(1, 3, figsize=(12, 4))
    print(f"  panellar shakli: {axes.shape} (1D — axes[i])")

    print("\n=== 2. axes[i] (1D indeks) ===")
    axes[0].hist(np.random.normal(0, 1, 100))
    axes[1].hist(np.random.normal(5, 2, 100))
    axes[2].hist(np.random.normal(-3, 1, 100))
    print("  3 panel: axes[0], axes[1], axes[2]")

    print("\n=== 3. Sarlavhalar ===")
    for i, nom in enumerate(["A", "B", "C"]):
        axes[i].set_title(nom)
    print(f"  sarlavhalar: {[axes[i].get_title() for i in range(3)]}")

    print("\n=== 4. Tushuntirish ===")
    print("  (1, 3) — 1D massiv (axes[i], 2D emas)")
    plt.close(fig)
    print("  ⭐ 1 qator — 1D indeks (axes[i])")


if __name__ == "__main__":
    main()

Natijaning muhim qismi:

=== 1. subplots(1, 3) — 1D ===
  panellar shakli: (3,) (1D — axes[i])

=== 2. axes[i] (1D indeks) ===
  3 panel: axes[0], axes[1], axes[2]

=== 3. Sarlavhalar ===
  sarlavhalar: ['A', 'B', 'C']

=== 4. Tushuntirish ===
  (1, 3) — 1D massiv (axes[i], 2D emas)
  ⭐ 1 qator — 1D indeks (axes[i])

Nima ko'rsatdi: 2.2-bo'lim.

Misol 3 — sharey (taqqoslash)

"""sharey: umumiy o'q (real matplotlib)."""

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np


def main() -> None:
    np.random.seed(0)
    a = np.random.normal(50, 5, 100)
    b = np.random.normal(50, 20, 100)   # kengroq tarqoq

    print("=== 1. sharey=True (umumiy y) ===")
    fig, axes = plt.subplots(1, 2, sharey=True, figsize=(10, 4))
    axes[0].hist(a); axes[1].hist(b)
    y0 = axes[0].get_ylim()
    y1 = axes[1].get_ylim()
    print(f"  panel 0 y: {tuple(round(v) for v in y0)}")
    print(f"  panel 1 y: {tuple(round(v) for v in y1)}")

    print("\n=== 2. Umumiy o'q (bir xil) ===")
    print(f"  y bir xil: {y0 == y1}")

    print("\n=== 3. Nega muhim ===")
    print("  turli y — taqqoslash noto'g'ri (5.13 aldash)")

    print("\n=== 4. Tushuntirish ===")
    print("  sharey — bir xil masshtab (to'g'ri taqqoslash)")
    plt.close(fig)
    print("  ⭐ sharey — umumiy o'q (taqqoslash)")


if __name__ == "__main__":
    main()

Natijaning muhim qismi:

=== 1. sharey=True (umumiy y) ===
  panel 0 y: (0, 20)
  panel 1 y: (0, 20)

=== 2. Umumiy o'q (bir xil) ===
  y bir xil: True

=== 3. Nega muhim ===
  turli y — taqqoslash noto'g'ri (5.13 aldash)

=== 4. Tushuntirish ===
  sharey — bir xil masshtab (to'g'ri taqqoslash)
  ⭐ sharey — umumiy o'q (taqqoslash)

Nima ko'rsatdi: 2.4-bo'lim.

Misol 4 — Seaborn subplots (ax=)

"""Seaborn subplots: ax= (real seaborn/matplotlib)."""

import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns


def main() -> None:
    np.random.seed(0)
    df = pd.DataFrame({
        "maydon": np.random.normal(80, 20, 100),
        "narx": np.random.normal(240, 60, 100),
        "yosh": np.random.normal(15, 5, 100),
    })

    print("=== 1. subplots + Seaborn ax= ===")
    fig, axes = plt.subplots(1, 3, figsize=(12, 4))
    sns.histplot(df["maydon"], ax=axes[0])
    sns.histplot(df["narx"], ax=axes[1])
    sns.histplot(df["yosh"], ax=axes[2])
    print("  3 Seaborn histplot — 3 panel (ax=)")

    print("\n=== 2. Har panel (ax=axes[i]) ===")
    print(f"  panellar: {len(axes)}")

    print("\n=== 3. Axes-daraja ===")
    print("  histplot — ax= (panelga); displot — o'z subplots'i")

    print("\n=== 4. Tushuntirish ===")
    print("  Seaborn (chizadi) + Matplotlib subplots (to'r)")
    fig.tight_layout()
    plt.close(fig)
    print("  ⭐ Seaborn subplots — ax= (axes-daraja)")


if __name__ == "__main__":
    main()

Natijaning muhim qismi:

=== 1. subplots + Seaborn ax= ===
  3 Seaborn histplot — 3 panel (ax=)

=== 2. Har panel (ax=axes[i]) ===
  panellar: 3

=== 3. Axes-daraja ===
  histplot — ax= (panelga); displot — o'z subplots'i

=== 4. Tushuntirish ===
  Seaborn (chizadi) + Matplotlib subplots (to'r)
  ⭐ Seaborn subplots — ax= (axes-daraja)

Nima ko'rsatdi: 2.5-bo'lim.

5. To'g'ri va noto'g'ri tushunishlar

6. Keng tarqalgan xatolar va yechimlari

1. figsize kichik

plt.subplots(3, 3)   # standart 6x4 (siqiladi)                    # ⚠️
plt.subplots(3, 3, figsize=(12, 12))   # katta                    # ✅

2. tight_layout unut

fig, axes = plt.subplots(2, 2)   # panellar to'sadi               # ⚠️
fig.tight_layout()   # joy                                         # ✅

3. axes indeks (1D/2D)

fig, axes = plt.subplots(1, 3); axes[0, 0]   # 1D (xato)          # ⚠️
axes[0]   # 1D (to'g'ri)                                          # ✅

4. turli o'q (taqqoslash)

plt.subplots(1, 2)   # turli y (taqqoslash xato)                 # ⚠️
plt.subplots(1, 2, sharey=True)   # umumiy y                     # ✅

5. figure-daraja Seaborn

sns.displot(data, ax=axes[0])   # displot — ax= yo'q (xato)      # ⚠️
sns.histplot(data, ax=axes[0])   # axes-daraja                   # ✅

6. suptitle vs title

axes[0].set_title("Umumiy")   # bir panel (butun emas)           # ⚠️
fig.suptitle("Umumiy")   # butun figura                          # ✅

7. bo'sh panel

plt.subplots(2, 3)   # 6 panel, 5 grafik → 1 bo'sh               # ⚠️
fig.delaxes(axes[1, 2])   # bo'sh panelni o'chir                 # ✅

7. Integratsiya — bu bilim qayerda kerak bo'ladi

• 5.1-dars (o'tilgan): Figure/axes (bir panel)
• 5.7-dars (o'tilgan): Seaborn (ax=)
• 5.13-dars: Yolg'on grafiklar (sharey — masshtab)
• 5.14-dars: Amaliy loyiha (dashboard)
• 9-qism (reja): EDA (ko'p xususiyat)

8. Eng yaxshi amaliyotlar

1. subplots(qator, ustun) — panel to'ri.

2. figsize — katta (ko'p panel).

3. tight_layout() — joy (to'silish).

4. axes[i, j] (2D) / axes[i] (1D).

5. sharey=True — taqqoslash.

6. suptitle — butun figura.

7. Seaborn — ax= (axes-daraja).

8. fig.savefig — butun figura.

9. Amaliy topshiriq

Vazifa 1: Bashorat qiling

1.  # subplots nima?
2.  # fig va axes?
3.  # (2,2) nechta panel?
4.  # axes[i,j] nima?
5.  # (1,3) indeks?
6.  # figsize nima?
7.  # tight_layout?
8.  # sharey nima?
9.  # suptitle vs title?
10. # Seaborn ax=?
11. # displot ax=?
12. # nega subplots?

Vazifa 2: Xatolarni tuzating

1.  plt.subplots(3, 3)   # siqiladi

2.  fig, axes = plt.subplots(2, 2)   # to'sadi

3.  axes = plt.subplots(1, 3); axes[0, 0]

4.  plt.subplots(1, 2)   # taqqoslash

5.  sns.displot(data, ax=axes[0])

1.  figsize=(12, 12)

2.  fig.tight_layout()

3.  axes[0] (1D)

4.  sharey=True

5.  sns.histplot(data, ax=axes[0])

Vazifa 3: subplots

Modellang:

1. Panel to'ri
2. fig, axes
3. Ko'p grafik
4. Taqqoslash

Vazifa 4: axes

Modellang:

1. axes[i, j] (2D)
2. axes[i] (1D)
3. Har panel
4. ax metodlari

Vazifa 5: Joy va o'q

Modellang:

1. figsize
2. tight_layout
3. sharey
4. suptitle

Vazifa 6: Integratsiya

Modellang:

1. Figure/axes (5.1)
2. Seaborn ax=
3. EDA
4. Dashboard

Vazifa 7: O'ylash

Subplots bir figurada ko'p grafik beradi — taqqoslash uchun kuchli. Lekin sharey bo'lmasa, panellar turli masshtabda bo'ladi va bir xil balandlik turli qiymatni anglatishi mumkin (aldamchi taqqoslash). Nima uchun "bir figurada taqqoslash" masshtab masalasi tufayli xavfli bo'lishi mumkin, va Data Scientist subplots taqqoslashda nimaga e'tibor berishi kerak?

Nimani mustahkamlaydi: 2.4, 2.7-bo'limlar.

Xulosa

Bu darsda subplots'ni o'rgandik.

Eng muhim uch fikr:

1. subplots va axes. plt.subplots(qator, ustun) — panel to'ri (bir figurada ko'p grafik; fig butun figura, axes panellar massivi; 5.1 figure/axes ustida). axes indekslash — axes[i, j] (2D — (2,2)) yoki axes[i] (1D — (1,3)); har panel alohida ax (plot/hist/scatter/set_title — 5.1-5.6).

2. O'lcham, joy, o'q. figsize=(k, b) (umumiy o'lcham — ko'p panel katta; siqilish), fig.tight_layout() (panel orasi joy — to'silish), sharex/sharey (umumiy o'q — taqqoslash; turli masshtab aldaydi — 5.13), fig.suptitle (butun figura sarlavha — ax.set_title bir panel).

3. Seaborn va taqqoslash. Seaborn subplots — ax=axes[i, j] (axes-daraja: histplot, boxplot, scatterplot; figure-daraja displot/relplot/catplot — ax= yo'q, o'z subplots'i). Subplots — bir figurada ko'p grafik (taqqoslash, EDA — panel to'ri). Tuzoqlar: figsize kichik (siqilish), tight_layout unut (to'silish), axes 1D/2D (indeks), turli o'q (taqqoslash — sharey), figure-daraja Seaborn (ax= yo'q), bo'sh panel (delaxes).

Keyingi darsda rang, uslub va bezashni o'rganamiz: grafikni chiroyli va tushunarli qilish — rang xaritasi, uslub (style), yorliq, izoh (annotate), grid (5.1 bezash chuqurroq).